1. Field of the Invention
The present invention relates to motion detection, and more particularly, to a method and a related apparatus for determining a luminance threshold value of a video region. As a result, the luminance threshold value can further be utilized for establishing a background model of the video region, and can therefore be a criterion for motion detection.
2. Description of the Prior Art
Motion detection is a technique that determines whether moving objects exist in a video sequence and then tracks the moving objects, analyzing motion of the moving objects by means of detecting difference of video frames. This technique is widely applied in many fields, such as professional sport games, security systems and military surveillances.
In the conventional art, a method for detecting motion called background model method is popular. This method establishes a background model in the video sequence in advance and then separates the moving object from the video frame. The background model method firstly divides each frame into a fixed number of video regions, and generates a luminance threshold value of each video region. The generated luminance threshold values of the video regions are utilized as criterions for detecting the moving object(s) in the video sequence. Thus, when a total luminance value of the video region changes (e.g. the total luminance value of this video region is different between two contiguous video frames), and a difference between the change of the total luminance value and the luminance threshold value is higher than a certain amount, it is considered that there is a moving object existing in/passing through this video region. Then, by analyzing all the video regions that are judged as having a moving object existing in/passing through, the moving object can be picked up from the video sequence. The most often seen method for generating the luminance threshold value is that which samples the total luminance value of the video region in a certain period, and then obtains a relative steady level of the total luminance value in this period. This level of the total luminance value is utilized as the luminance threshold value within the period and as a criterion for determining whether a moving object exists in/passes through the video region. To obtain the luminance threshold value, the total luminance is sampled in this period, and after the period ends, all sampling results are averaged for representing the level of the total luminance value during the period. To exactly show the changes of the background model, this luminance threshold value needs to be updated often so that motion detection can be correctly performed when the background changes. The process regarding the aforementioned description is already summarized into steps 110-114 shown in FIG. 1. A detailed description is omitted here since these contents should be well known to those skilled in the art.
Hardware costs are considerable in the conventional art, wherein most hardware costs relate to storage apparatus used for recording each sampled result. For example, the luminance threshold value of a video region is attempted to be updated every 10 seconds, and the total luminance value is sampled 30 times in a second (that is, in a video device capable of generating 30 frames per second, the total luminance value of the video region is sampled for each frame). As for each video region, 300 storage spaces (e.g. register file) are needed for recording each sampled result, and also, an averaging circuit is necessary for generating the luminance threshold value of the video region. In addition, most of the hardware costs are consumed by the storage apparatus since the precision of the average calculation results from the sampling time and sampled data. If the sensitivity of the motion detection needs to be enhanced, the hardware costs of the conventional method will increase considerably.